Generative AI as a customer service management tool
Two years after Generative AI first hit the headlines, it feels like the right time to take a step back and think about the less obvious, but potentially more impactful, ways the technology can help customer service.
A lot of the investment in Generative AI for customer service has so-far been focused on tools to automate tasks carried out by Customer Service Representatives (CSRs). This includes chatbots and voicebots that interact directly with customers, as well as tools which help CSRs to take notes, look up information, and develop their skills. There is good logic to this focus: in most customer service operations, 70-80% of the cost is for frontline contact centre teams, and because Generative AI models offer a conversational interface, it makes sense that they should help with tasks that involve having conversations with customers.
With the benefit of using Generative AI, we have learned that it excels at a number of tasks beyond just having conversations:
Synthesis and summarization of key points from large-scale unstructured data
Simulation of human responses to information provided
These tasks overlap with some of the skills needed to be an effective customer service managers, leading to the question…
What should an AI-empowered customer service manager have at their fingertips?
Here are three management activities that could be enhanced by the capabilities of Generative AI
1. Knowledge curation
One of the most common use cases for Generative AI in customer service is to use Retrieval Augmented Generation (RAG) to provide dynamic knowledge content to CSRs based on the conversation being had with the customer.
This works well, until you have to deal with knowledge content that is incomplete, badly formatted, or just plain wrong. This is unfortunately the norm rather than the exception for customer service operations.
Generative AI can help with the curation of knowledge content too. For example, by analysing all knowledge articles and identifying those that give contradictory or ambiguous advice.
2. Workforce management
Automated forecasting, planning and scheduling of customer service workforce is more of a classic Machine Learning task than something suited to Generative AI. However, the conversational interface and human simulation capabilities of Generative AI can be put to use to improve speed and effectiveness of WFM teams. Imagine a conversational interface where a workforce planner can enter prompts like
Why were calls higher than expected yesterday and should we plan to increase capacity over the next few days?
or
Based on our contract terms with three different outsourcers, which site should we send excess volume to, to achieve the best value for money?
or
Based on previous responses, which team members are most likely to be willing to work an additional weekend shift this month?
An AI tool that can answer these types of questions could enhance the value that WFM can create.
3. Performance management
There are many Generative AI-based tools that offer one-on-one coaching and guidance to individual CSRs, but imagine being able to do this at scale across a whole contact centre operation. We will soon start to see tools that enable managers to see a “heat map” of performance across their whole team, to a fine level of detail. An example insight could be
Here is a group of 50 CSRs who struggle to explain one-off fees to customers and would benefit from additional training on this topic
Used wisely, these tools will enable managers to focus their time on driving performance in the priority areas where they can have the most impact.
Recommended news articles
Latest perspectives from BCG
European consumers are optimistic but still not spending
People in Europe feel more optimistic about their personal situation than about national issues, which has affected their spending practices, according to new BCG research on consumer sentiments, shopping habits, and preferences in the region.
We found that across the continent, 73% of people reported higher prices for goods and services in the first half of 2024, including for basics such as groceries, as well as for non-essentials, such discretionary items as clothes and toys.
During the same time, salaries and savings stayed the same or dropped, leading people to compensate by spending less on non-essentials.